If an LED is blinking through the hardware and the trigger is deactivated,
the driver will set the brightness to 0 to disable blinking. However, if
the LED driver doesn't support non-blocking operations, the request for
setting brightness to 0 will get queued. Then if a user requests a new
non-zero brightness before the request for 0 brightness is executed, the
LED blinking will not get disabled since the brightness will simply be
updated.
Fix this by flushing the workqueue before updating the requested
brightness.

Fixes: 4d71a4a12b13 ("leds: Add support for setting brightness in a synchronous way")
Signed-off-by: Eddie James <eajames@linux.ibm.com>
---
 drivers/leds/led-class.c | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/drivers/leds/led-class.c b/drivers/leds/led-class.c
index 6a8ea94834fa..861a1cb5df5b 100644
--- a/drivers/leds/led-class.c
+++ b/drivers/leds/led-class.c
@@ -55,6 +55,8 @@ static ssize_t brightness_store(struct device *dev,
 
 	if (state == LED_OFF)
 		led_trigger_remove(led_cdev);
+	/* flush out any request to disable blinking */
+	flush_work(&led_cdev->set_brightness_work);
 	led_set_brightness(led_cdev, state);
 	flush_work(&led_cdev->set_brightness_work);
 
-- 
2.27.0

On 5/20/22 12:11, Eddie James wrote:
> If an LED is blinking through the hardware and the trigger is deactivated,
> the driver will set the brightness to 0 to disable blinking. However, if
> the LED driver doesn't support non-blocking operations, the request for
> setting brightness to 0 will get queued. Then if a user requests a new
> non-zero brightness before the request for 0 brightness is executed, the
> LED blinking will not get disabled since the brightness will simply be
> updated.
> Fix this by flushing the workqueue before updating the requested
> brightness.


I just realized this never made it anywhere. Pavel, any thoughts?


Thanks,

Eddie


>
> Fixes: 4d71a4a12b13 ("leds: Add support for setting brightness in a synchronous way")
> Signed-off-by: Eddie James <eajames@linux.ibm.com>
> ---
>   drivers/leds/led-class.c | 2 ++
>   1 file changed, 2 insertions(+)
>
> diff --git a/drivers/leds/led-class.c b/drivers/leds/led-class.c
> index 6a8ea94834fa..861a1cb5df5b 100644
> --- a/drivers/leds/led-class.c
> +++ b/drivers/leds/led-class.c
> @@ -55,6 +55,8 @@ static ssize_t brightness_store(struct device *dev,
>   
>   	if (state == LED_OFF)
>   		led_trigger_remove(led_cdev);
> +	/* flush out any request to disable blinking */
> +	flush_work(&led_cdev->set_brightness_work);
>   	led_set_brightness(led_cdev, state);
>   	flush_work(&led_cdev->set_brightness_work);
>

Apologies for any chaos this may cause, I'm replying from Gmail since
this doesn't appear to be in my inbox.

On Wed, 5 Apr 2023 at 17:26, Eddie James <eajames@linux.ibm.com> wrote:
>
>
> On 5/20/22 12:11, Eddie James wrote:
> > If an LED is blinking through the hardware and the trigger is deactivated,
> > the driver will set the brightness to 0 to disable blinking. However, if
> > the LED driver doesn't support non-blocking operations, the request for
> > setting brightness to 0 will get queued. Then if a user requests a new
> > non-zero brightness before the request for 0 brightness is executed, the
> > LED blinking will not get disabled since the brightness will simply be
> > updated.
> > Fix this by flushing the workqueue before updating the requested
> > brightness.
>
>
> I just realized this never made it anywhere. Pavel, any thoughts?

Submit a [RESEND] please.

Also, please ensure you re-run get_maintainer.pl before doing so.

> > Fixes: 4d71a4a12b13 ("leds: Add support for setting brightness in a synchronous way")
> > Signed-off-by: Eddie James <eajames@linux.ibm.com>
> > ---
> >   drivers/leds/led-class.c | 2 ++
> >   1 file changed, 2 insertions(+)
> >
> > diff --git a/drivers/leds/led-class.c b/drivers/leds/led-class.c
> > index 6a8ea94834fa..861a1cb5df5b 100644
> > --- a/drivers/leds/led-class.c
> > +++ b/drivers/leds/led-class.c
> > @@ -55,6 +55,8 @@ static ssize_t brightness_store(struct device *dev,
> >
> >       if (state == LED_OFF)
> >               led_trigger_remove(led_cdev);
> > +     /* flush out any request to disable blinking */
> > +     flush_work(&led_cdev->set_brightness_work);
> >       led_set_brightness(led_cdev, state);
> >       flush_work(&led_cdev->set_brightness_work);
> >

Any update on this?


Thanks,

Eddie


On 5/20/22 12:11, Eddie James wrote:
> If an LED is blinking through the hardware and the trigger is deactivated,
> the driver will set the brightness to 0 to disable blinking. However, if
> the LED driver doesn't support non-blocking operations, the request for
> setting brightness to 0 will get queued. Then if a user requests a new
> non-zero brightness before the request for 0 brightness is executed, the
> LED blinking will not get disabled since the brightness will simply be
> updated.
> Fix this by flushing the workqueue before updating the requested
> brightness.
>
> Fixes: 4d71a4a12b13 ("leds: Add support for setting brightness in a synchronous way")
> Signed-off-by: Eddie James <eajames@linux.ibm.com>
> ---
>   drivers/leds/led-class.c | 2 ++
>   1 file changed, 2 insertions(+)
>
> diff --git a/drivers/leds/led-class.c b/drivers/leds/led-class.c
> index 6a8ea94834fa..861a1cb5df5b 100644
> --- a/drivers/leds/led-class.c
> +++ b/drivers/leds/led-class.c
> @@ -55,6 +55,8 @@ static ssize_t brightness_store(struct device *dev,
>   
>   	if (state == LED_OFF)
>   		led_trigger_remove(led_cdev);
> +	/* flush out any request to disable blinking */
> +	flush_work(&led_cdev->set_brightness_work);
>   	led_set_brightness(led_cdev, state);
>   	flush_work(&led_cdev->set_brightness_work);
>